Case Study: Major Global Bank achieves real-time fraud detection with StarTree

Major Global Bank Uses StarTree for Real-Time Fraud Detection

Major Global Bank, a New York City–based payments platform provider, needed a faster way to detect and stop fraud across more than 70 million transactions annually. Its existing stack struggled with low throughput and high costs because transactional systems like MySQL and Bigtable were being pushed beyond their OLTP limits for real-time analytical fraud monitoring.

To solve this, Major Global Bank implemented StarTree, powered by Apache Pinot, alongside Confluent for real-time data streaming and CDC from MySQL. With StarTree, the bank improved end-to-end latency by 16x, reduced p99 query latency from 300 ms to 50 ms, and achieved p50 query times as low as 7 ms, while supporting over 300,000 queries per day and up to 1,200 QPS.